Some NA Z-scores from findTrends #32
Description
Hello,
I was following along the example analysis with some real Visium HD data, and noticed that after running findTrends, some Z-scores were NA, which caused correctZBonferroni and vizColocDotplot to fail. I ran 5 permutations, and fortunately at least 1 permutation had a non-null value for each reference-neighbor pair. After running meltResultsList, I figured I could just filter out rows with null Z-scores, but wondered if this throws off the multiple-testing calculation. Also, perhaps NAs should not be present in the first place.
One abnormality about my data is that I randomly downsampled it just to speed up interactive testing, but the cell-type factor variable has all the levels. Aside from maybe having lower cell density than typical, I don't believe there should be issues with this data.
Let me know if I can provide more info.
Best,
-Nick